Re: stty 0; stty 19200 on modem1, 1 patch...

[itschere@TechFak.Uni-Bielefeld.DE]

Salut Thierry!

> Apart from that, I also have a (buffered and interrupt-based) device
> driver for the Centronics port, if someone is interested...

 I would be quite interested, not only because I could then stop
finishing my own driver... :-)

 Since completely integrating new device driver into the kernel is
probably a longish thing to do, I was wondering if it wasn't a good
idea to fix the biosfs so that it would be possible to replace an
internal bios device with a driver with the same name. Just to be
transparent...

 Also I was wondering where the consensus should go how these driver
report incoming characters to the kernel. As far as I see, your
modm0dev uses a daemon which periodically checks for characters.
Which is the only reason I don't have installed it all the time,
cause it really messes debug output a lot :-(

 My idea was that perhaps the so far used variable 'kintr` should be
turned into a 32 bit vektor and made accessible through the kernel
pointer, so that every device driver gets a bit for its own. Would
of course need some numbering of device driver, limit them to not more
than 32 at a time and perhaps be somehow slow if checking if done in
the context switch routine, if you want a quick response, but...

 Ah yes, and then I'm known to be not severly interested in not
breaking existing code sometimes... ;-)

 Point is that I think this should perhaps allow a quick integration
into the kernel, in spite of rewriting most of it.

au revoir,
TeSche
-- 
PS: If the above written looks weird, then that's probably because it _IS_.
Torsten Scherer (Schiller, TeSche...), University of Bielefeld, Germany
==> itschere@techfak.uni-bielefeld.de / tesche@hrz.uni-bielefeld.de <==